Douglas Aberdeen
Papers
2
Total Citations
144
H-Index
2
About
Douglas Aberdeen is a leading researcher in reinforcement learning, with a particular focus on policy-gradient methods for partially observable Markov decision processes (POMDPs). His foundational work, including the highly cited 2003 paper "Policy-Gradient Algorithms for Partially Observable Markov Decision Processes" (75 citations), addresses the challenge of learning optimal behaviors in complex, real-world environments where agents have incomplete information—such as robot navigation, speech recognition, and stock trading. Aberdeen's major contribution lies in developing scalable internal-state policy-gradient algorithms that enable effective learning even when memory is required, overcoming a key limitation of earlier methods. His 2002 paper on this topic (69 citations) introduced innovations that improved the performance of POMDP solvers in memory-demanding tasks. With over 140 combined citations for these two seminal works, Aberdeen's research has had a lasting impact on the field, providing both theoretical insights and practical algorithms that continue to influence modern reinforcement learning and autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2Scalable Internal-State Policy-Gradient Methods for POMDPs69 citations · 2002